A computational psychological approach to complex emotion: interest links complex emotion and attractiveness

While emotions in daily life are often complex-involving multiple, sometimes contradictory feelings-most emotion research has used simplified laboratory stimuli designed to elicit single emotions. We conducted a descriptive study documenting emotional complexity in naturalistic emotional experiences. Sixty participants rated their emotional responses to 261 film clips (brief segments within 60 s) immediately after viewing, assessing 13 discrete emotions, valence, understanding, and attractiveness (15,660 observations total). Four patterns emerged: (1) Emotional complexity was prevalent (78% of clips elicited multiple emotions, M = 2.8 when > 1 reported). (2) Complexity strongly correlated with attractiveness (r = 0.79), with clips eliciting more diverse emotions being rated as more attractive. (3) Interest showed distinctive properties, statistically mediating relationships between 11 of 12 other emotions and attractiveness with uniform strength (b-path CV = 3.2%); its association with attractiveness remained robust after controlling for clip comprehensibility. (4) Substantial individual differences existed in complexity sensitivity (range: 8-85% of clips) without affecting mean attractiveness. We propose the Interest-Complexity Framework: a three-stage descriptive model wherein diverse emotions are elicited (Stage 1), interest evaluates and integrates complexity (Stage 2), and integrated experiences inform attractiveness judgments (Stage 3). While correlational data cannot establish causality, this framework generates testable experimental predictions. Our systematic documentation of emotion co-occurrence patterns in aesthetic stimuli provides foundational data for affective computing, media psychology, and emotion AI, demonstrating that complex emotions are the norm rather than exception in naturalistic contexts.